Impact of Changes to LIS Status on Part D Drug Utilization March 20, 2012

1. Impact of Changes to LIS Status on Part D Drug UtilizationMarch 20, 2012 Thomas Kornfield, Health Insurance Specialist Medicare Plan Payment Group Centers for Medicare and Medicaid Services (ARS Response Card: Channel 41)

2. Disclosure “I, Tom Kornfield, declare no conflicts of interest or financial interests in any product or service mentioned in this presentation, including grants, employment, gifts, stock holdings, or honoraria.”

3. Introduction and Overview • The Part D benefit provides low-income beneficiaries with reduced or eliminated cost sharing • Individuals who change low income status may choose to alter drug prescription utilization due to changes in cost sharing • Purpose of study is to evaluate how changes in cost sharing can affect prescription drug utilization

4. Learning Objectives • Analyze changes in prescription drug utilization based on low income status • Determine which drug class has the highest expenditures for low income beneficiaries

5. Low Income Cost Sharing Categories for 2010

6. Methodology • Methodology • Use data from Integrated Data Repository (IDR) at CMS to identify beneficiaries who had: • Low income in 2009, not in 2010 • Low income in 2010, not in 2009 • Enrollment Requirements • Enrolled in a plan providing Part D coverage in each month from January 2009 to December 2010 • Enrolledin the same plan throughout the year • Did not change address during the year • Low Income Requirements • 12 months of continuous LI status in either 2009 or 2010 • PDE claims were pulled from the IDR

7. Study Population Characteristics

8. Characteristics of Study Population, Low Income 2009 not 2010

9. Characteristics of Study Population by Plan Type in Non LI Year

10. Study Population Characteristics – Distribution by Gender and Age, Low Income 2009 but not 2010

11. Study Population Characteristics – Distribution by Gender and Age, Low Income 2010 but not 2009

12. Study Population Characteristics – Distribution by Age and Presence of Drug Claim, Low Income 2009 but not 2010

13. Study Population Characteristics – Distribution by Age and Plan Type in 2009, Low Income 2009 but not 2010

14. Distribution of Low Income Beneficiaries with Status Change

15. LI Beneficiaries With Status Change as Percent of All LI Beneficiaries Connecticut

16. Distribution of Plan Switchers by Group

17. Number of Prescriptions for Brand Drugs, Top 10 Drug Classes by 2009 Spending, Low Income in 2009 and not 2010

18. Total Spending for Brand Drugs, Top 10 Drug Classes by 2009 Spending, Low Income in 2009 and not 2010

19. Total Spending for Top Ten Drug Classes Based on 2009 Spending, Low Income in 2009 and not 2010

20. Total Spending for Top Ten Drug Classes Based on 2010 Spending, Low Income in 2010 and not 2009

21. Differences in Spending and Risk Score, Low Income 2009 and not 2010

22. Differences in Spending and Risk Score, Low Income 2010 and not 2009

23. Changes in Spending and Risk Scores by Original Reason for Medicare Entitlement, Low Income 2009 not 2010

24. Changes in Spending and Risk Scores by Gender, Low Income 2009 and not 2010

25. Changes in Total Spending by Plan Type, Low Income 2009 and not 2010

26. Distribution of Total Spending by Year, Low Income in 2009 and not 2010 Beneficiaries with Less than $10,000 Spending

27. Distribution of Spending, Low Income in 2009 and not 2010

28. Distribution of Beneficiaries with Changes in Spending from 2009 to 2010, Low Income in 2009 and not 2010

29. Distribution of Total Spending by Year, Low Income in 2010 and not 2009 Beneficiaries with Less than $10,000 Spending

30. Potential for Future Research • Conduct Regression Analysis • Dependent variable would be change in total drug costs between non LI year and LI year, • Independent variables are age, Part D risk score, Medicaid status, census region, plan type • Conduct regression separately for the following groups: • Low income in 2009, not in 2010 • Low income in 2010, not in 2009 • Could also predict likelihood that beneficiary will use a brand name prescription drug • Create cohorts for other years

31. Conclusions • Small number of individuals changed LI status from 2009 to 2010 • More people lost than gained LI status from 2009 to 2010 • Loss of LI status leads to lower total prescription drug spending and less brand prescription drug use • Future research could analyze how changes in spending are affected by other characteristics

32. Assessments

33. Assessment Question 1 How much higher are prescription drug expenditures for individuals in the year after they obtain low income status? No difference Less than 1 percent higher 1 to 10 percent higher More than 10 percent higher None of the above

34. Assessment Question 2 Which drug class has the highest expenditures for low income beneficiaries in 2010? Blood Sugar Lowering Drugs (Diabetes Drugs) Triglyceride and Cholesterol Lowering Drugs (Hyperlipidemia) High Blood Pressure (Hypertension) Anti-Dementia Drugs None of the above

35. Questions?

36. Contact Information For more information please contact: Thomas Kornfield Medicare Plan Payment Group (410) 786-3116 thomas.kornfield3@cms.hhs.gov

37. Presentation EvaluationPlease get your ARS Response Card ready